As we noted last week, Spain has forced the hand of other Eurozone bank regulators by declaring it will release the results of recent ECB stress tests, which earlier were to be published only on an aggregated basis, not bank by bank.
There is still a good bit of confusion as to what happens next. The flurry of announcements by other eurozone leaders is that they will also release the results of stress tests, but it is clear that these will be new stress tests, not the ones already completed (note this story by Reuters, for instance, which refers to past as well as prospective stress test results).
We said we did not expect this movie to end well, even though the euro rallied impressively on the news. First, the data was to be released only on an aggregated basis before precisely because French and German authorities had not liked the idea of showing bank by bank results. Not only are French and German banks rather heavily exposed to Club Med sovereign debt, but they are likely to have dodgy exposures from the 2007-2008 crisis that are still marked very generously (even though the US is the land of “extend and pretend”, the eurozone banks have marked down less of their dodgy debt than their US peers).
Those concerns are already being reflected in the marketplace. From the Financial Times:
Fears are rising over French and German banks’ exposure to weaker economies in the eurozone such as Greece, Portugal and Spain after moves to publish bank stress tests in Europe.
Investors warn the tests could expose the European banking system’s interdependence and spread contagion, which started with Greece, to the continent’s two biggest economies.
Elisabeth Afseth, fixed-income strategist at Evolution, said: “The stress tests have focused some investors’ minds on the exposure of France and Germany to the peripheral economies. This could put the French and German bond markets under pressure.”
The Bank for International Settlements published figures last week showing that French and German banks were particularly exposed to Greece, Ireland, Portugal and Spain…..The BIS said German and French banks had combined exposures of $958bn to Greece, Ireland, Portugal and Spain at the end of 2009 in its quarterly report published on June 14.
So we have problem one sorta confirmed, that publishing credible stress test results will make some banks look bad, which is precisely what the authorities want to avoid (hold that thought, we will come back to it in due course).
But recall….the reason the Spanish (and now the eurozone members who have been forced to fall into line) wanted the stress tests to be published was to restore confidence in their banks.
The problem was that confidence created by the stress tests was not simply the result of publishing a report card with As, or at least nothing worse than a B-, but having a credible process (at least as far as the not-always discerning media, analysts, and investors were concerned).
John Hempton provided a useful reminder last week (hat tip Richard Smith):
The Scandinavian banking crisis was solved in the following manner
(a). The banks were guaranteed
(b). Someone independent of the banks was invited in to reassess bank capital.
(c). The banks were then told how much capital they had to raise. They had a fixed period of time to raise it.
(d). If they could raise it – well and good – and they kept operating. If they could not the Government injected capital cancelling existing equity as it went and where it could ultimately wind up with 100 percent ownership. They were not afraid of the “n-word” (ie nationalisation).
The American solution worked almost identically except for step (d). In America
(a). The government told us that there would be “no more Lehmans” and they kept telling us and giving banks access to additional funds until we all knew the banks were effectively guaranteed,
(b). They had a “stress test” to assess how much capital to raise.
(c). The banks were told how much capital to raise and given a time. They raised it in common equity.
(d). If the banks could not raise the capital the government injected capital as common shares until they had enough. Note that the US process could not wind up with 100 percent ownership of a bank – and the “n-word” was not used. If the US had run on the Scandianvian formula Citigroup would be entirely government property.
Yves here. Hempton leaves out a few key steps in the US process, namely a well orchestrated PR campaign to talk up the banks, plus lots of theater during the stress test process, which enhanced perceptions that the tests were tough (here the propensity of bankers to fight tooth and nail over every perceived indignity was a huge plus).
The part that Hempton finesses is that the “no “n” word in the US meant the stress tests were inadequate, but that was very adeptly camouflaged. As we detailed on this blog ad nauseum while the tests were on, they were a joke. There was no effort to validate the banks’ data (by contrast, normal protocol for a troubled bank is to sample loan files). The banks were asked to run various scenarios on their own risk models, the very ones that had performed so well during the crisis. The exam was also skewed towards the lending side of the business, when the banks all had large, and potentially deadly, toxic assets and problematic liabilities via derivatives in their trading operations.
Moreover, the “adverse” scenario in the stress tests as far the performance of the economy was concerned, was far too optimistic. And it was also vastly too generous in its treatment of second mortgages exposures. The stress tests called for assuming only 13-14% losses (except for Citi, which was told to assume 20%). As Mike Konczal pointed out this year, House Financial Services Committee chairman Barney Frank has deemed seconds to have “no economic value”. The four biggest banks hold $477 billion of seconds. Even assuming, charitably, as Konczal did, that they are worth somewhere between 40 and 60 cents on the dollar, that means losses of between $190 billion and $285 billion. That in turn implies that the banks have hole in their balance sheets roughly $150 billion larger than what the stress tests showed, and that in turns implies their capital raising was $150 billion short of what was necessary.
Oh, and these banks were all allowed to pay back TARP funds because they were supposedly healthy. And the big reason for their anxiety to escape the TARP? Not because it was good for their enterprise, but to free themselves of executive pay restrictions.
So why did the stress tests work? Team Obama is great at propaganda. The media fell completely into line. One indicator: about midway through the two-month process, Bill Black appeared on the widely watched Bill Moyers show and pronounced them a sham. Not a single journalist contacted Black about the stress tests. USA Today did ask him to run an op-ed on the stress tests, then turned it down when he took his usual candid stance.
Of course, impressive-looking first and second quarter 2009 earnings didn’t hurt (leading bank analyst Meredith Whitney deemed them to be “manufactured” but also conceded the banks could keep it up for a bit) as did concerted squeezes of short interests in bank stocks.
So let us consider the obstacles to the Europeans conducting “successful” (meaning convincing to the market) stress tests:
1. Lack of a mechanism to credibly inject serious amounts of capital should that be shown to be needed. Revealingly, Merkel assured the press that the banks could rely on the eurozone rescue facilities, the same ones that are already seen as inadequate to deal with looming sovereign debt problems. As John Hempton noted:
This solution [the Scandanavian style stress test process] works provided you have sovereign solvency. A sovereign can do this if it can print money (I will go into mechanics later) but cannot do it on a gold-standard or Euro standard. The Scandinavians needed to de-peg their currency from Europe to achieve their solution and to this day there is a Swedish, Norwegian and Danish Kroner. Finland alone took up the Euro – but Finland has no large domestically owned banks.
The solution will work in Europe too provided the currency of the PIGS is separated from the Euro. It will not work otherwise because step (a) above – the Government guarantee of the banks – is not possible.
2. The various banking regulators will need to agree on the scenarios. One of the reasons that the US tests charade worked was because Treasury and the Fed had worked closely together during the crisis (remember, Geithner had just come over from the New York Fed). So with the exception of possible complications coming from Shiela Bair at the FDIC, they key players were working together and even if they had private differences, always presented a unified front in public. By contrast, every step of the eurozone crisis has shown friction and conflict, with agreement being reached painfully, at the 11th hours.
Banking regulators being less colorful and public sorts, the disarray may not be visible, but the most likely result instead is “adverse” scenarios so generous as to be unconvincing to the markets.
3. Adequate disclosure needs to be made of the scenarios and the bank-level results. Again, if national bank regulators worry that their banks will be exposed as being weak, they will fight for less disclosure than the market will deem adequate. This would put the eurozone officials back where they started, perhaps even worse off, since the failure to reveal enough will be seen as confirmation that there is indeed something to hide.
The problem at its root appears to be that the Spanish wanted to release the stress test results to prove that their banks are adequately capitalized. Even if that is true, it is certain not to be true for all the major Eurobanks. Eurozone officials, lacking a credible mechanism to shore up weak bank, have recklessly committed to showing the dirty laundry of weak banks, or more likely, engage in finesses which will simply confirm, if not heighten, existing doubts.
Update: FT Alphaville is also on a stress test kick today, see “Some stresstestimates” and ”